Forecasting Ski Demand: Comparing Learning Curve and Varying Parameter Coefficient Approaches

Demand for skiing expanded rapidly in the 1980s, fell quite dramatically at the start of the 1990s as the economy declined but has not subsequently recovered. Two possible explanations are explored. The first is based on perceiving skiing as a new product to most consumers, which reached maximum growth in 1989. Current levels now largely represent ‘repeat buyers’. The alternative approach sees the growth as the result of economic factors, particularly credit conditions. The importance of these factors was not, however, constant, and grew with the changes in the financial system. Thus the recovery had a muted eect. These two approaches are modelled, estimated and the results compared by both residual and ex post forecasting analysis. The paper concludes that the varying coecient econometric model probably produces the most reliable forecasts. Copyright # 1999 John Wiley & Sons, Ltd. KEY WORDS forecasting; skiing; tourism; time series; learning curve; econometric; varying coecien